JPH01157686A - Apparatus and method for video scrambling based upon space filling diagram - Google Patents

Abstract

PURPOSE: To operate the device within the frequency band of a standard video signal by deciding a pseudo random space filling curve(SFC) providing a sequence mapping adjacent pixels so as to obtain higher coding gradient. CONSTITUTION: The sequence of pixels is converted by changing scanning sequence without changing pixel values. Thus, a frame buffer 22 is used to store grating graph display of adjacent pixels in horizontal and vertical directions of a video image. Since the storage of correlation between adjacent pixels is desired, a scanning pattern being a consecutive curve is selected and it is called a space filling curve(SFC). The scanning pattern is given through address designation arranging again the retrieval of the digital sequence of the frame buffer 22 and a control computer 24. Thus, a high coding gradient is obtained the device is operated within the frequency band of a standard video signal.

Description

TECHNICAL FIELD OF THE INVENTION The present invention relates to techniques and apparatus for scrambling video, and more particularly to scrambling and scrambled digital representations of video signals before and after a video transmission procedure. The present invention relates to a video scrambling technique that uses a pseudo-random space-filling curve to undo (hereinafter referred to as descrambling).

[Background of the invention]

Video scrambling techniques are useful for home cable applications where it is desired to limit programming to authorized subscriber viewers.
It is used in television and is also known to be used in military applications such as transmitting video signals from airplanes.
Ru. Between these two extremes of use are industrial applications such as video conferencing, video telephony, two-way television, etc.

The difficulty in encoding a video signal is that the signal is analog and has a very high transmission rate, and reordering the analog video information onto one line is difficult because the signal and noise spectrum at the receiver is very high. With the advent of VLSI electronics technology, which has only become practical in the past few years, very fast digital storage methods have been introduced that allow video information to be stored in frame buffers in real time. Speed b that can be memorized and retrieved
-Can now be used using RAM. This has led to the possibility of scrambling video signals by digital methods without touching the transmission format.

Examples of these digital methods include line reversal, line position modulation, line segment swapping, and line distribution.
All of these are R,K
``Highly Secure Television Transmission Using Digital Processing'' by Upniki) et al.

IEEE Publication 84CH2069-3 (1984). This paper describes how to widen the frequency band of the initial video signal so that it can be used by standard video receivers. draws attention to the need to choose a suitable scrambling algorithm to prevent
7) As described in U.S. Patent No. 4.070,693.
Ru.

Examining these scrambling techniques from the perspective of encoding strength, correlation techniques compare lines or segments until the first sequence of lines or segments is found in order to recover a significant portion of the entire image. It will be appreciated that matching allows the synopsis to be de(11) coded.

An additional digital method of scrambling a video signal is described in Taylor British Patent No. 1.59.
No. 0,579 requires a random permutation of pixels within a line. It is stated here that the property of a video scrambling device is to destroy the relationship between adjacent image points, resulting in a higher transmission bandwidth requirement for the video signal. Well, the distortion of the image makes the technique ineffective.

One very popular video scrambling device is provided by the method known as Video Cipher, offered on the market by M/A-COM Linkabit Corporation. Cipher Happy
``Video Cipher Scrambling System'' by (P, Mulroney) et al., I
EEE Publication 84CH2069-3 (1984). This system is also described in European patent application no.
However, since the line portion is long, the number of jumps in the video signal is kept low and therefore increasing the frequency band is not important.

However, long line sections facilitate establishing correlation in the scrambling algorithm by comparing and matching line sections.

It would therefore be desirable to provide a video scrambling technique that has higher encoding efficiency and operates within the frequency band of standard video signals.

[Summary of the invention]

It is therefore an object of the present invention to encode a video signal using a pseudo-random space-filling curve with a high (random scanning pattern) for scanning the pixel addresses of a video image stored in a frame buffer. The purpose of this method is to provide a method that can be used with video scramblers and descramblers, and is characterized by greater coding strength than conventional techniques (13). The subject is the establishment of simple correlations by comparing and matching line parts (
・. In addition, this method is characterized by narrowing the video signal frequency bandwidth for typical images.
・Ru.

In accordance with a preferred embodiment of the present invention, a method for encoding a video signal is provided, the method comprising the following steps.

storing pixel values of some video information lines representing at least some of the video pixels included in the video signal in positions in the frame buffer; - Random vertical and horizontal movement of frames to determine a pseudo-random space-filling curve that gives a sequence corresponding to the position of the frame buffer, and encoded frames for the transmission of the video signal in the manner Frame (14
] ・Step to specify the address of the Baafa position.

In a preferred embodiment, the method of the present invention is applied to the operation of a video scrambling device that uses a frame puncher to store a grid graph representation of horizontally and vertically adjacent pixels in a video image. There is. A space-filling curve (SFC) is defined as a Hamiltonian bus that passes through each pixel exactly once. The SFC is generated by a scrambler to take advantage of the two-dimensional nature of the video image displayed by the original source of the video signal. Using pseudo-random SFC to define the sequence in which frame buffer addresses are read provides a highly irregular scanning pattern of adjacent pixels of the video image. This is because there is an extremely large probability of an irregular path based on horizontal and vertical movement between the pixels.

The encoding provided by the video scrambler in the transmission of the video signal is decoded by the descrambler. The legs of both the scrambler and descrambler
By using a keystream that is shared between
The latter generates the same pseudo-random space-filling curve as the scrambler, thus providing the descrambler by calling the correct location of the frame buffer to decode the original video signal.

Pseudo-random SFC is generated by each execution of an algorithm whose output is in the domain of a set of Hamiltonian paths in a rectangular lattice graph given by a frame buffer. - Once sFc is selected from the set of Hamiltonian paths, the addresses of the frame buffers are used to define the sequence in which they are read.

After a predetermined time, the first SFC is replaced by another bus from the family of Hamilton passes, further increasing the security of the encoding method.

A feature of the method of the invention is that the frequency band of the video signal is narrow for normal typical images. This is because SFC provides a scan pattern that preserves the correlation between adjacent pixels. The vertical and horizontal broken lines that characterize SFC are sequential frames. Traditionally performed over a group of buffer locations, a raster scan creates a more "local" scanning pattern. Make it smaller (・ru.

Another embodiment of the present invention is that performing the descrambling does not have to produce an SFC (.
Ru.

In yet another embodiment, the SFC of the
・ru) is selected according to the encoded signal included in the transmission signal.

Another feature of the invention is the application of SFC to sequential consecutive frames of video images. Storing video information in a sequential frame memory (17) allows three-dimensional SFC to be used, since the sequential possible positions are separated by successive video frames. This further increases the security of the encoding method and further reduces the frequency band of the signal due to the 3-D locality of the scan.

Other features and advantages of the method of encoding according to the invention will become apparent from the accompanying drawings and the description below.

For a better understanding of the invention with respect to embodiments of the invention, reference is made to the accompanying drawings and appendices, in which numerals in the same drawing indicate corresponding elements throughout the drawings. .

[Detailed description of preferred embodiments]

Referring now to FIG. 1, there is shown an overall (7) block diagram of a direct broadcast satellite system incorporating a preferred embodiment of the method of the present invention. The need to securely transmit video information in such systems has many applications, including TVROs (TV receiving stations, holding video conferences, videophones, analog facsimile, and military communications systems) in cable TV systems. As will be discussed further herein, the present invention provides a method that is much more powerful in encoding than conventional methods of scrambling video information.

In the system of FIG. 1, scrambler 10 receives a standard video signal that is encoded and transmitted via uplink transmitter 12 to communication satellite 14.

Reception of the transmitted signal is by a satellite receiver 16, and a descrambler 18 decodes the encoded video signal and restores it to its original form as output. The inputs and outputs of scrambler 10 and descrambler 18 are both video signals, typically in a standard format, and are therefore fully compatible.

When performing the encoding method of the present invention, the scrambler 10
interferes only with the order of the video signal, not with the shape of the signal. The scrambler 1o and the descrambler 18 are modules that are added to the conventional system, while other components are modified (・(19)). between the cent and video signal cable inputs (・
It is envisioned as a "black box", which typically comes from a receive-only antenna. Cable or telephone lines can be replaced by satellite communications links.

Now, referring to FIG. 2, the scrambler 10 of FIG.
A system block diagram is shown illustrating the encoding method of the present invention as used in.

The encoding method is based on rearranging the permutation of transmitted pixels of the video signal.An analog/digital (A/D) converter 20 converts the video signal into a digital signal.
The video signal information is then sent to frame buffer 22 for storage in raster scan format. Addressing and control computer 24 allows pixels to be stored in frame buffer 22.
The order in which such searches are performed is controlled. A digital-to-analog (D/A) converter 26 converts the permuted digital sequence back to an already scrambled analog video signal and transmits the scrambled video to the uplink transmitter 12 or other Subject to transmission by a transmission station.

According to the encoding method of the invention (.), the scrambled video signal provided by the scrambler 10 must satisfy seemingly contradictory requirements (.. on the one hand, it is possible to attack correlation-based attacks). We need to have some randomness to survive (・.On the other hand,
Its smoothness requires an autocorrelation function equal to or even higher than the autocorrelation function of the first raster scan (without being affected as the frequency band increases). do.

As is well known, conventional raster scanning involves scanning pixels from top to bottom and from left to right within a frame. The method of the present invention is based on transforming the pixel order by changing the scanning order without changing the pill cell values. To this end, the frame buffer 22 is Since it is desired to preserve the correlation between adjacent pixels, a scan pattern that is a continuous curve is chosen. This is called a space-filling curve (SFC). This type of curve takes advantage of the similarities between vertical autocorrelation and horizontal autocorrelation between pixels to solve the two-dimensional nature of a video image. The scanning pattern is
It is provided by an addressing and control computer 24 which reorganizes the retrieval of digital sequences in the frame buffer 22.

Referring now to FIG. 3, the method of the invention includes different horizontal 0 . ,) A space-filling curve (SFC) is shown that provides a scanning pattern that allows vertical progression between pixels on a video line. The scans must be continuous in the image (i.e. pixels that are adjacent in the scan are also adjacent in the image, and the scrambled video signal remains smooth.The space-filling curve (S
FC) is a Milton pass, which visits every pixel of the image and never intersects itself, so there are no jumps in the scan, and no second transmission of the same pixel.
. Using a pseudo-random SFC to define the sequence in which frame buffer addresses are read gives a highly irregular scan-stagger scanning pattern of adjacent pixels in a video image. This is because there is a possibility of obtaining irregular paths based on horizontal and vertical movement between pixels.

According to the present invention, the encoding method pseudo-randomly selects the SFC and also selects the SFC frame buffer 2 at
The selection of the SFC used as a scanning pattern to reorganize the search for sequences of pixels within the 2σ '1'' SFC follows the mathematical treatment of the representational geometry of what is currently being presented in the video image. Based on the implementation of an algorithm that efficiently generates SFC from a set of Hamiltonian paths.

Is the combination of vertices in G an integer coordinate? A lattice graph is an infinite graph consisting of all points of a plane with two vertices connected if and only if the (Euclidean geometry) distance between them is 1. It is an induced sub-graph. Let Vx and Vy k be the coordinates of the apex ■. The rectangular lattice graph R(m, n) is a lattice graph, and the combination of its vertex angles is (■: 1≦Vx≦m and 1 ≦vy≦n).Thus, if the frame buffer 22 can be thought of as a rectangular grid graph whose vertices are pixels, then the SFC is R(m, n) and θ
This becomes a Hamilton pass.

The problem of finding Hamiltonian paths for specific vertices of rectangular lattice graphs was discussed in the paper ``Hamiltonian paths in lattice graphs'', A. Itai et al., Siam, J., Computer (1982). This paper shows how to create one such path.According to the present invention, many such paths can be created efficiently. For further discussion, see S. Even, "Combinatorics of Albo-Rhythms," Manoku Millan, New York (1973).

(The algorithm that efficiently creates all the Hamiltonian paths and "habo" Hamiltonian paths of a 24-rectangular lattice graph R(m, n) is called the Ham algorithm, which can be defined as follows. Note that the algorithm appears there in a form suitable for coding in the Pascal programming language.

Algorithm RAM” Initialization: 1←1 Repetition: Execute the following while i<(m-2).

1, degree (■, one = 0 a(・dani, right (Vl one and downward (
Select V-type for connection. (All selections are inevitably made from the leftmost vertex) 2.V last +V.

3. If V exists, do the following.

(B) If degree (Vr) is 20, connect ■1 to the left (■1) and downward (Vr) (inevitable choice) tb+If degree (vr) = 1, next (・So Gyokudan left (Vr
]=Partner (Vr) or ■r2V last, ■1
Connect downward (Vr).

(Small ℃・Do not close the circle (・necessary choice) For connection with Vr, randomly left (Vr) and downward (Vr)
Noa(・Choose the crab.

(This is due to randomness in circuit construction. (This is where you enter) 4.1　←1+1 Connect the final two paths to each other to form a wide-area circuit.

Processing: Connection 1, create a path to return to ■, and downward (■1) through the leftmost vertex of row (m-1).

2. The rightmost vertex of row (m)? then Vk and downward (V
Make a passage between (9 on the right) and return.

(20-) The rightmost end point is now downwards (Vk)”
It is k-x. ) 3. Make a passage that crosses the lower (Vk) and lower (Vk-1) at and d.

(2s i 4- to -2 4 (the three rightmost endpoints are now V1+□5 below (v
, +□] and ■1) i+1〆1, then do the following.

If the talum humidity ■1+ and the downward direction (■1+1) are partners: fjlVi+□ and ■1 (7) A(・Dani, create a direct passage.

Strike down (V, +□) and down (■1), make a direct passage.

There is (・If V1+, and ■1 are partners: (1) Right (■) and v1+□A(・da (including these?)) Through V′ for V′ with the function ■1+ , and downward (v1+
,] Noa~・Dano (Make a return passage.

(11) Create a return passage (on the right) between ■1 and below (■1) via the left (v').

Otherwise, if (■1 + 0. Downward (v, +□ru and ■1
is the end point of a different passage) Choose at random from the above direct passage or return passage.

(2υ (The two rightmost endpoints are now lower (V1) and Vl-1.) Create a crossing path between +b + lower (V) and lower (V, -1).

(C1i 4-i -2 5, stop (If 1-4-1=1, the cross passage is downward (V
i) so that a wide area circuit is achieved. ) Some definitions are required to understand the algorithm.

First, degree (V) is defined as the number of things that are close to the previously selected ■ in the algorithm. Second, the vertex ■, and V2I'! , are defined as partners if they are (then) the endpoints of the same path.

In a certain row, 5■1 is the leftmost vertex V, and degrees (V) < 2
, and let ■ be the degree (V) < 2 at the leftmost vertex. There is a vertex whose degree is <2 (・, then there is V
It can be thought of as (・. When thinking about the vertex ■, the right (V)
, left (V), lower (V, ), and upper (V) are respectively to its left.

These are the vertices to its right, below it, and above it. Vertex■1.
When considering , this means the vertex of the i-th column and third row.

For simplicity, after each step in the algorithm, a variable ■□
, ■ r (degree (V) and partner (V)) are defined as above. Justifying this assumption is that updating these variables can be done in small steps as shown below. The terms ``select a neighbor j'' and ``connect vertices'' () are used interchangeably, and the particular vertices will be clear from the text.

The following update rules apply:

1. Initially, every vertex is its own partner and has degree O.

2. Initially, ■ in each column, and Vr are the leftmost and rightmost vertices in the column.

3. After connecting V□ and ■2, K+ partner (V) and partner (■2) become partners.

(29) tbt degree (■,) and degree (■2) are increased by 1.

(C1 If vl was ■, then new ・■1 is degree (right (V,)) (respectively, right (■,) or right (right ・(■□) ].

tdl If ■1 is ■□, the new ・Vr is degree (left (Vl)) (when 2 or = 2, respectively)
Left (V□) or left (left (■□)).

If in step 3 (C) ■2 = right (■□) or degrees (
It was noted that when ■2) is exactly at 1 before it is connected to ■1, degree (right (Vl)) = 2 (・
. In this case, ■1 - right (■2). Because 5 degrees (
This is because the right (■□〕) is 2. Similarly, step 3 (d) Te+! , ■ - Left (■□) degree (■2 When the degree was 1 before being connected to ■□, exactly dust (left [■□]) = 2.

Also, if ■ most i (algorithm Ham's suup2
If [as determined by
I have ■1 Hana (・It got a lot of attention・.

Algorithm Ham uses frame buffer 22Q
``Habo''Hamilton's circuit is constructed by scanning the track from top to bottom.

A convention that can be used is to count rows from top to bottom and columns from left to right. Algorithm σ, ;(
i-i') After repeating the fragrance, rectangle R(i-1,
1) is satisfied by setting all endpoints in the first row to advance the endpoints. There are no endpoints left inside R(1,n), and no path becomes a closed circuit (
・. This is accomplished by having the endpoints of each path identify each other as partners. In the last two lines, the )s are connected to each other to obtain a wide area circuit.

The results shown in FIG.
C: Yes. These movements can be indicated by left, right, upward or downward. Only sequences in the left, right, upward or downward direction are required by the moving address control 24. This sequence is defined by the SFC (given 31). Partial comparison and matching prevents attacks in the system.Using SFC allows for a reduction in memory footprint, allowing for example memory containing 25x25 pixels to be used, in this case 2300 pixels. There are several possible SFCs.

Other possibilities exist for two-dimensional SFCs involving diagonally moving permissive scan patterns to form paths in triangular or honeycomb graphs σC. The Hamiltonian path principle holds true in these cases, following appropriate algorithms to achieve them. Instead,
A large number of consecutive curves can be used as they fill the required space in the beginning.

Referring now to FIG. 4, a schematic block diagram of a video scrambler 10 employing the method of the present invention is shown. Scrambler 10 includes a sync separator 32 for detection of horizontal and vertical sync signals that are provided to control logic 34.
A video signal 30 is supplied to. After synchronous separation, it is applied to an analog/digital converter 20. The video signal 38, which is digital, is applied to one of the frame stores 42 or 44 (forming the frame buffer 22faO) via respective lines 46 or 48 to a frame switch 40 controlling a Q signal supply. The particular frame store is selected by control logic 34 depending on whether it stores an odd or even number of fields (.Q fields).

In the preferred embodiment, two frame stores 42 and 44 are used.
are used, and these allow each to operate slower than if one frame store were used. This means that one video signal can be written into one while the other is read, so that if the TV picture has 60 frames per second, each of the 5 frame stores 42 and 44 will contain that video signal. It takes 1/60 second to complete writing or reading according to the operation of the frame switch 40 and the frame switch 50 on the output side.

It will be appreciated that alternative embodiments may use more than one frame store, each of which may also have a slower response than those described above, without affecting the overall system response. will be done.

The two frame stores 42 and 44 are fast (video RA)
M, each organized into a matrix of e.g. 256x256x8 bits, which is used to store the incoming video on lines 46 and 48, all control signals and addresses are routed via control bus 54 to control logic logic 34. Provided to frame stores 42 and 44 by write address generator 52 under supervision.

According to the method of encoding of the invention, the key stream signal 5
6 is provided to the blank/pull controller 58 for processing using the Ham algorithm in the preferred embodiment to generate a pseudo-random vacancy-filling curve. The keystream signal 56 may be provided as the output of a pseudo-random number generator initiated by a "seed" number.
The outputs are left, right, up, and down direction commands (L, Fl
, U, D, ) and these instructions are sent to the read address generator 60. The direction command is translated by read address generator 60 to a corresponding physical address in frame stores 42 and 44. These physical addresses are stored in one of a pair of read address lookup tables 62 and 64.

If two or more frame stores are provided, the number of read address look-up tables will be correspondingly large.

In operation, when one of the frame stores 42 and 44 is full and the other is switched in to help with video, data from the first frame store is read address-lookup. Tables 62 and 640) are read using addresses from the appropriate one. These addresses are read (35) and sent to address bus 68 via address logic 66.
is supplied to As a result, the digital video signals are read from the frame stores 42 and 44 in a different order than they were written to, thus causing the switching of the frame switch 50 to be controlled by the control logic 34. Accordingly, scrambled video signals 70 and 72 are provided.

The scrambled video signals 70 and 72 are provided to a digital to analog converter 26Vc for conversion to an analog video signal 79, 5 and an output mixer 76 which receives the key stream 56. The audio portion of the video signal, which can be encoded by digitization and scrambling, is processed by an audio processing unit 80 and provided as an audio signal 81 to an output mixer 70. In this way, a complete analog video signal output 84 is generated by two video signals 79. Key stream signal 56, audio signal 81, and synchronization generator 8
6 and a periodic signal 85 (36).

In FIG. 5, a schematic block diagram of the descrambler 18 is shown. Descrambler 18 provides video signal 90 to sync separator 92 for separating horizontal and vertical sync signals that are provided to control logic 94 . After synchronous separation,
The video signal 90 is analog/analog for conversion to digital form.
A digital converter 96 is provided. Digital video signal 98 is provided to a frame switch 100 which controls the provision of signals to one of frame stores 102 or 104 via respective lines 106 or 108. This particular frame store is selected by control logic 94 for odd or even field storage.

The two frame stores 102 and 104 are high speed video R
AM, each organized into a 256X256X8 bit matrix (same as scrambler 10), line 10
It is for storing videos on 6 and 108. All control and address signals are routed through control bus 115 to write address logic 110 under the supervision of control logic (37 94).
and write address lookup table 112
and 114 frame memory fit 102 and 104
is supplied to

According to the encoding method of the invention, the keystream signal 56 used in the scrambler 10 is provided to the descrambling controller 1]6 for processing to generate a pseudo-random space-filling curve using the same Ham+ algorithm. The keystream signal is also the same “
The numbers can be generated by a pseudo-random number generator with a "seed". In such cases, the key stream need not be provided to mixer 76 (. Instead, a synchronization signal may be used. A sequence of downward direction commands (L, R, U, D) that is provided to write address generator 118. Direction commands are provided by write address generator 118 to the physical A physical address is a pair of write address, look,
is stored in one of the up tables 112 and 114.

In operation, the incoming video signal 9o is passed to the scrambler 10 by pixel relocation means according to the SFC.
It has already been scrambled.

In a descrambling system, unit pixels are written to frame stores 102 and 104 using addresses created by the same algorithm using the same keystream signal 56 used in scrambler 10. At this time, the key stream signal 56 is separated (or locally generated) within the key separation device 120. These addresses are accessed via write address lookup tables 112 and 114 to address f.
f119J1 ] 5 to the write address logic 110. As a result, the digital video signals will be written into each of frame stores 102 and 104 in the same order as they were read by scrambler 10, thus decoding the first image. This causes the descrambled video signal 122 to be re-retrieved into raster scan format by turning off the operation of the frame switch 125 controlled by the control logic 94.
and give 124.

In response to operation of the read address generator 126 via the address bus 127, the descrambled video signals 122 and 124 are sent to a digital-to-analog converter 128 for conversion to an analog video signal and to an output mixer. 132. The audio portion of the video signal 90 is encoded by the scrambler 10 (
- Tara is requested to be decoded, processed by an audio gross sensor 134 and provided as an audio signal 135 to an output mixer 132. In this way, the fully analog video signal output 13G
video signal], 30, an audio signal 135, and a synchronization signal 138 from a synchronization generator 140.

As shown in the preferred embodiment, the keystream signal 5
6 is transmitted together with the video signal. The method of the present invention is based on SF
Since it contains an efficient algorithm for generating C, this signal will often be sent to initiate the generation of another SFC, thereby increasing the encoding power of the C system. The keystream 56 itself may be encoded, or in another embodiment, the keystream may be coded with a code number based on the use of a "smart card." In this configuration, the descrambler combines the use of a smart card with a transmitted "seed" number, which together generate a keystream signal 56 via a pseudorandom number generator.

Another embodiment of the invention is a transmission of SFC, so the descrambler does not have to generate SFC (the transmission is sent along with the video signal. In yet another embodiment,
Several SFCs are stored in the descrambler prior to its operation, and the appropriate SFC (used in the scrambler 10) is used to decode the encoded signal contained in the transmission. The keystream 56 can be used for the purpose of selecting the appropriate SFC for the descrambler 18 or can be used to encode the transmission of the SFC itself. It would be possible to reduce the cost of the descrambler.

As mentioned above, it is a special feature of the method of the present invention that the frequency bandwidth of the video signal is continuously reduced but never increased for typical images. is true because the SFC provides a scanning pattern that preserves the correlation between adjacent pixels. is a "local" scan pattern for a group of frame buffer locations in the sequence, and this can reduce the amount of pixel-to-pixel variation and narrow the frequency band.

In FIG. 6, this narrowed frequency band characteristic is represented as a graph illustrating the operation of an autocorrelation applied to a scrambled video signal produced by the method of the invention. Curve A shows a typical raster scan, where the autocorrelation is low (... because the pixel values are
This is because as the scan travels from one side of the image to the other as fast as possible, it changes at the same rate. In contrast, curves B and G show two scans using two different SFCs, and also exhibit thicker autocorrelation between adjacent pixels. Curve D shows low autocorrelation for random scanning, where no relationship between pixels is obtained.

FIG. 7 is a graph comparing the normalized integral of power spectral density produced by a scrambled video signal with that of conventional raster scanning. The frequency bandwidth of the four scans (total frequency band = 100%) is proven by showing the integral of the power spectral density. Therefore, the larger the integral, the thicker the scan for frequency band purposes.The straight line is that of a random scan.The raster scan A is thicker.
・The power spectral density belongs to the two upper kerfs B and C, which are thicker than that of the random scan.
Corresponds to scanning using .

Another feature of the invention is the application of SFC to a consecutive series of frames of video images. Storing video information in a series of frame stores allows two-dimensional SFC to be used. because,
This is because the series of possible positions are contained within a cubic grid constructed by successive video frames. This further increases the security of the encoding method. Because it offers extra degrees of freedom in three dimensions, the number of possible SFCs is much larger (2710,000 possible for a three-dimensional SFC where each frame's storage contains 250x250 pixels). There are some changes.

(44) To understand our approach to three-dimensional SFC generation, we will briefly explain another method for creating two-dimensional SFCs. This method is called the “tiling and convergence method,” and initially consists of rectangular grids. Based on covering all the vertices of the graph by the initial circuits that are separated and then creating a Milton pass (by merging these circuits with any Hamilton path) .For this method, some definitions are required.

Let G be a directed (・na(・graph). Separate circuits C□, C that cover all vertices of G
2. Consider Ck. If two of these circuits, Gi and CJ, have edges e1=(V□, v□′) in C1, and edges e2=[■2・■2′)( =■21.■2〕〕If the side f is (V1', V2'), it is defined as an adjacent circuit. f□ and f2 are e□ and C
2 into a square (・.If Gi
If and Cj are adjacent circuits, then they are connected to one line Ci3 covering exactly the same vertices as C1 and GJ (45) by replacing e□ and C2 by fl and f2. This replacement is a legal replacement and changes (e□, 02] to (C1, 0])
Representation as a legally replaced couple (LRC) in the... attracted attention. The feet and legs of the two circuits (・There are several legitimate replacements, but who among them?
It was noted that the following would be chosen at random (・.

In this way, the number of circuits is reduced by one with a single legitimate replacement. Thus, if a process begins with a circuit, after k-1 legal permutations one Milton circuit remains.

This leads directly to the following definition and algorithm for finding random Hamiltonian circuits in general lattice graphs.

Each of k divided circuits C1, C2,...
・.

If Ck is given in G, then Vk=(cm.

When C2,...,Ck) and gk=((c,,cko): C1 and CJ are adjacent circuits), the graph G2(
Vk.

Define Ek). For each (Ci, CJ)←Ek (
46] allows the following relation to be obtained for (L): (L, j)k= ((el, e2) : e□ and 821
! (LBC in (C,,Cj)) It has been noted that in the lattice graph, the edges of the couple that are legitimately replaced are adjacent edges located in different circuits (. If 2 If two circuits (・displacement are also surrounded by others (・none), then the two circuits are named as being outside.Circuit C1
, C2, . . . , Ck are considered to be tiles of G. If all the circuits are outside in pairs,
Those circuits are named external circuits and the tiles are named external tiles.

Now consider a three-dimensional S F' C, which can be said to be an extension of the three-dimensional picture in which the basic circuit is a cube of eight vertices, merging with two-dimensional tiles. is here called Gu, and Cu is a node with six adjacent cuboids in Gk - the six cubes are flat and lIh with respect to Gu at the surface. The algorithm provided here for creating a 3D SFC span-cube includes the following steps.

1° Fit several parallel squares into a cube with 8 vertices.

2. Randomly select a tree T that acts as a bridge to GkVC.

3. Less than 6 degrees (・Choose a knotty root of T
Turn all sides toward the leaf.

4T is scanned using BFS, the method of searching for width strips.

"Assignment" (see below) processing in each visited section Cu? Use this to allocate one circuit l- to one in Cu(G) to perform a legal replacement.

“Assignment J processing is: 1 Randomly choose one of the four possible cubic circuits.

Depending on the shape of the 2Cu, the entrance edge is determined by the front edge and the edge leaving the cube circuit is chosen to each outgoing edge.

3. Inform all children of each person leaving.

Each node enters through one edge, which will be called the entry edge, and leaves via the 0-5 edge, which will be called the leaving edge on the left. In Figures 8a to 8d, the artificial edge is shown as being the upper left side of the cube. For a cube (a cube that can be rotated around an axis of symmetry) with the label one (・teshi・na(・a cube that can be rotated around an axis of symmetry) there is exactly one circuit that covers the vertices of the cube. ] For a hanging cube (one fixed in space) of the hanging label, there are four possible covering cube circuits as shown in Figures 8a to 8d.

In the "assignment" process, the edge e1 toward and away from the node Cu in the TO is selected, and the match is relayed to Cu. The side in C that is close to e is necessarily 1 of Gu.
u will be the entrance side. The entry node of the root is chosen at random from the side with an edge in T (・, which is the reason for choosing the root with degree less than 6. Nodes in T go outside up to 5 There are 32 possible shapes for each node, which can have edges.For each cubic circuit, each outgoing edge (TO
There is at least one way to assign a leaving edge (in G) to a group. Explicitly (whenever an assignment is made, one of the possible assignments must be chosen at random). This leads to the formal definition of the assignment process as shown above. '', ``node'', ``father'', and ``child''' are all attached to the bridging tree T (・ru. All edges are assumed to be in G (・ro. Assignment It was recalled that the processing should be carried out when visiting the node Cu in T(11).
U's father has already visited (・ro.

Beginning (Algorithm, in Space Cube) C
u is the root of T, and the entrance side of the root is as described above.

This three-dimensional method of creating SFC also has the advantage of making the width of the frequency band even smaller, since SFC makes use of frame-to-frame correlation, i.e. Because there is a high correlation between two pixels at the same spatial location, there is redundancy from frame to frame. Three-dimensional SFC In this way, 200 pixels (N
Two pixels apart (2009) are three-dimensional science fiction
When C is used, consecutive frames are separated by a Euclidean distance of about 20-30 pixels (N) corresponding to three-dimensional blocks of the frame store. Where two-dimensional SFC is used, two pixels separated by 200 pixels are approximately 50-60 pixels apart in the corresponding image.
Pixels (No. 75) will be separated. In raster scanning, 200 pixels apart in a video signal will be approximately the same in an image.

This comparison is about the smoothness of the video signal in three dimensions.

Therefore, the lower frequency bandwidth is indicated.

Referring to FIG. 9, the automatic correlation function of the raster scan signal (A), the random two-dimensional SFC scan signal, and the random three-dimensional SFC scan signal is shown. This shows an improvement in the performance of a two-dimensional SFC scan signal and a further improvement in the performance of a three-dimensional SFC scan signal compared to a conventional raster scan signal.

Although the invention has been described in connection with some particularly selected embodiments, it is not intended to be limited to what has been described and further modifications may suggest themselves to those skilled in the art. however, such modifications are intended to be within the scope of the claims.

[Brief explanation of the drawing]

FIG. 1 is an overall block diagram of a direct broadcast system incorporating a preferred embodiment of the encoding method of the present invention, and FIG.
FIG. 3 is a block diagram of a system illustrating the inventive encoding method used in the system of FIG.
4 is a schematic block diagram of a video scrambler using the method of the present invention shown in FIG. 2; FIG. 5 is a schematic block diagram of a video descrambler using the method of the present invention; Figure 6 is a graph showing the behavior of auto-correlation applied to the scrambled video signal generated by the method shown in Figure 2. Figure 7 is a graph showing the behavior of the automatic correlation applied to the scrambled video signal generated by the method shown in Figure 2. Graph 5 Comparing Conventional Raster Scanning of Integration and Random Pixel Order Exchange Figures 8a to 8d illustrate four possible cubic circuits for use in the method of the present invention in a three-dimensional SFC. Figure 9 is a diagram showing a comparison between the automatic correlation function of two-dimensional and three-dimensional SFC and conventional raster scanning.
18...Descrambler, 20...Analog/
digital converter, 22...frame buffer,
24゛°"Control computer, 26...Digital/analog converter, 30...Video signal (53) Figure 6, Figure 8 (A> (B)
(D>Kuu 7 Figure 0yoz5χ 50; l 75λ
DAηχ tail R diagram

Claims (1)

[Claims] 1. In a method of encoding a video signal, frame
storing pixel values of a set of video information lines representing at least a portion of a video image included in the video signal at locations in the buffer; a frame buffer mapping random directions between adjacent pixels of the portion of the video image; A pseudorandom space-filling curve (S
FC), and addressing a frame buffer location according to the determined sequence for reading pixel values for transmission of a video signal in an encoded state. A method of encoding video signals. 2. The method of claim 1, wherein the frame buffer forms an element of a video scrambler that generates the SFC and addresses the position of the frame buffer accordingly. A method of encoding a video signal with characteristics. 3. The method of claim 1, wherein the SFC converts almost every pixel of the video image into
A method of encoding a video signal, characterized in that it is a circular Hamiltonian path by moving between adjacent pixels. 4. The method according to claim 2, wherein the SFC is a Hamiltonian pass that cycles through almost all pixels of the video image by moving between adjacent pixels. How to encode. 5. The method of claim 1, wherein the SFC provides mapping of at least two of vertical, horizontal and diagonal movement between substantially all adjacent pixels of a video image portion. A method of encoding a video signal with characteristics. 6. The method according to claim 2, wherein the encoding provided by the video scrambler in the transmission of a video signal is performed by the descrambler using a shared keystream between the scrambler and the descrambler. The descrambler generates the same pseudo-random SFC used in said scrambler, thereby re-storing the video signal to the frame buffer location contained in the scrambler to re-store the original video signal. A method of encoding a video signal, characterized in that it provides a scrambler with the correct sequence to write. 7. The method of claim 3, wherein the encoding provided by the video scrambler in the transmission of a video signal is performed by the scrambler using a shared keystream between the scrambler and the descrambler. The descrambler uses the same pseudo-random S used in the scrambler.
1. A method of encoding a video signal, comprising providing a scrambler with the correct sequence for generating FC and thereby writing the video signal into the locations of a frame buffer included in the scrambler for restoring the original video signal. 8. The method of claim 4, wherein the encoding provided by the video scrambler in the transmission of a video signal is performed by the descrambler using a shared keystream between the scrambler and the descrambler. The descrambler generates the same pseudo-random SFC used in said scrambler, thereby re-storing the video signal to the frame buffer location contained in the scrambler to re-store the original video signal. A method of encoding a video signal, characterized in that it provides a scrambler with the correct sequence to write. 9. The method of claim 5, wherein the encoding provided by the video scrambler in the transmission of a video signal is performed by the descrambler using a shared keystream between the scrambler and the descrambler. The descrambler generates the same pseudo-random SFC used in said scrambler and thereby re-stores the video signal to the frame buffer locations contained in the descrambler to re-store the original video signal. A method of encoding a video signal, characterized in that it provides a scrambler with the correct sequence to write. 10. The method according to claim 2, wherein the encoding provided by the video scrambler in the transmission of a video signal comprises a plurality of encodings, the selection being made according to a signal indicating the SFC used in the scrambler. decoded by a descrambler storing in itself the SFC of the video signal, thereby providing the descrambler with the correct sequence to write the video signal to a frame buffer location included in the descrambler to re-store the original video signal. A method of encoding video signals. 11. The method according to claim 3, wherein the encoding provided by the video scrambler in the transmission of a video signal is selected according to a signal indicating the SFC used in the scrambler. decoded by a descrambler storing in itself the SFC of the video signal, thereby providing the descrambler with the correct sequence to write the video signal to a frame buffer location included in the descrambler to re-store the original video signal. A method of encoding video signals. 12. The method according to claim 4, wherein the encoding provided by the video scrambler in the transmission of a video signal comprises a plurality of encodings, the selection being made according to a signal indicating the SFC used in the scrambler. decoded by a descrambler storing in itself the SFC of the video signal, thereby providing the descrambler with the correct sequence to write the video signal to a frame buffer location included in the descrambler to re-store the original video signal. A method of encoding video signals. 13. The method according to claim 5, wherein the encoding provided by the video scrambler in the transmission of a video signal comprises a plurality of encodings, the selection being made according to a signal indicating the SFC used in the scrambler. decoded by a descrambler storing in itself the SFC of the video signal, thereby providing the descrambler with the correct sequence to write the video signal to a frame buffer location included in the descrambler to re-store the original video signal. A method of encoding video signals. 14. In the method according to claim 2, the SFC used for encoding in the video scrambler is included in the transmission of the video signal because it is used by the descrambler; Rambla is
A method of encoding a video signal, characterized in that a correct sequence of writing the video signal to the locations of a frame buffer included in a descrambler is provided for restoring the original video signal. 15. The method of claim 3, wherein the pseudorandom SFC is performed by executing an algorithm whose output domain is the set of Hamiltonian paths in a rectangular lattice graph given by the scrambler frame buffer. A method for encoding a video signal, wherein the SFC is selected from the set of Hamiltonian passes and used to determine the sequence in which frame buffer addresses of the scrambler are read. 16. In the method according to claim 15,
A method of encoding a video signal, characterized in that, after a predetermined time, the first SFC is replaced by another SFC from a group of Hamilton passes. 17. A method according to claim 1, characterized in that the frequency band of the video signal is reduced and the SFC preserves the correlation between adjacent pixels of the video image. How to make it. 18. The method of claim 1, wherein the SFC is applied to a sequence of successive frames in a video image to determine the number of possible frame buffers in a cubic grid formed by the successive video frames. A method for encoding a video signal, characterized in that it provides a three-dimensional sequence. 19. A video scrambling system comprising the method of claim 1.